On Tue, 28 Feb 2023 at 22:44, Helmut Grohne wrote:
> Would you agree to either include the PR in your upload or pass
> --debarch=$DEB_HOST_ARCH instead of --debarch=auto in debcrossgen?
Because of #1032168 the current priority is to get 1.0.1 to testing.
Thus I'm not going to do any changes to t
Hi Jussi,
On Mon, Feb 27, 2023 at 12:53:16AM +0200, Jussi Pakkanen wrote:
> I filed an MR upstream with all changes that should go in the next
> release (which is to say they won't make it to bookworm):
>
> https://github.com/mesonbuild/meson/pull/11464
This all looks good to me.
> I also uploa
On Sun, 26 Feb 2023 at 21:39, Helmut Grohne wrote:
> Given that, I think you can move forward with the change post bookworm.
> Thanks for bearing with me and sorry for the inadequate remarks.
I filed an MR upstream with all changes that should go in the next
release (which is to say they won't m
Hi Jussi,
On Sun, Feb 26, 2023 at 05:40:06PM +0200, Jussi Pakkanen wrote:
> I find this a bit confusing. As far as I can tell those are in the file:
>
> https://github.com/mesonbuild/meson/blob/master/mesonbuild/scripts/env2mfile.py#L134
Mea culpa. My look into it was too brief. you're entirely
On Sun, 26 Feb 2023 at 09:22, Helmut Grohne wrote:
> I also spent a brief look at the env2mfile implementation and am
> bewildered that it essentially discards the experience gained with
> debcrossgen over the years. The mapping replacements for armhf, i386 and
> mips64el are missing. Compiler fl
Hi Jussi,
On Sun, Feb 26, 2023 at 03:16:08AM +0200, Jussi Pakkanen wrote:
> I uploaded a new version to experimental that does exactly what I
> described earlier. Basically everything should "just work" without any
> changes needed. Feel free to try it out and report any failures to me.
I'd appre
On Sat, 25 Feb 2023 at 18:20, Helmut Grohne wrote:
> While I appreciate the idea, I think the result is not useful.
> debcrossgen is something that we want to delete before too long. It also
> is platform logic by definition as it is specific to Debian. Carrying
> platform logic (like touching D
Hi Jussi,
On Sat, Feb 25, 2023 at 02:04:32PM +0200, Jussi Pakkanen wrote:
> What I wanted was to isolate _all_ code that has to do with platform
> detection and selection in exactly one place. The only thing
> debcrossgen should do is to call a different command. It should have
> _zero_ "platform
On Sat, 25 Feb 2023 at 12:08, Helmut Grohne wrote:
> I think I understand this part. At no point did Simon or me ask for
> changing the interface of env2mfile. What we asked for is to retain the
> interface of debcrossgen, which is Debian-specific. These requirements
> seem quite easy to reconcil
Hi Jussi,
On Sat, Feb 25, 2023 at 03:18:57AM +0200, Jussi Pakkanen wrote:
> The reason for this is that env2file is meant to be general. It needs
> to support custom toolchains, distro package building and the like.
> Debcrossgen only does Debian packaging (and is quite a bit older,
> roo).
I thi
On Fri, 24 Feb 2023 at 23:29, Simon McVittie wrote:
> The option seems to be debcrossgen --arch rather than --debarch, although
> it becomes meson env2mfile --debarch.
The reason for this is that env2file is meant to be general. It needs
to support custom toolchains, distro package building and
On Fri, 24 Feb 2023 at 22:42:09 +0200, Jussi Pakkanen wrote:
> Using the log info I managed to find out why it does not work. The
> tests that I have specifically pass --debarch=xxx to debcrossgen which
> then passes it on to env2mfile and things work. However the package
> builds do not do that, i
On Fri, 24 Feb 2023 at 19:38, Helmut Grohne wrote:
>
> Could it be that you tested env2mfile, but not debcrossgen? It would
> always exit non-zero no matter what. No error message. This really
> should have been caught by the "crossbuild" autopkgtest, but wasn't. I
> don't quite understand why. I
Hi Jussi,
On Fri, Feb 24, 2023 at 02:27:20PM +0200, Jussi Pakkanen wrote:
> I do have a test that runs the new cross file generator. I ran it
> again by hand and it does work. The files it produces are not exactly
> the same, but the information within seems to be the same. The only
> major differ
On Fri, 24 Feb 2023 at 08:26, Helmut Grohne wrote:
> I think you know, but let's include the answer here for clarity: You
> change debcrossgen to become a wrapper of env2mfile. However doing so
> requires that env2mfile actually works and last time you attempted this,
> env2mfile was so horribly
Hi Jussi,
On Sun, Aug 28, 2022 at 10:57:00PM +0300, Jussi Pakkanen wrote:
> The first step would be to change the default packaging rules to stop
> using debcrossgen. This is now possible since there is a new enough
> Meson version in unstable. What's the correct way of getting that
> default chan
On Tue, 16 Aug 2022 at 12:51, Simon McVittie wrote:
> debcrossgen puts c_args, cpp_args, c_link_args and cpp_link_args in the
> [properties] section of the cross-file that it generates, but recent
> versions of Meson want this to be in the [built-in options] section
> instead.
We have been movin
Package: meson
Version: 0.63.1-1
Severity: minor
debcrossgen puts c_args, cpp_args, c_link_args and cpp_link_args in the
[properties] section of the cross-file that it generates, but recent
versions of Meson want this to be in the [built-in options] section
instead.
18 matches
Mail list logo